The New Method / Substep / Concept Idea Thread

[DoAnKhang]

Its like using TSLE to force twisted TTLL but my method uses VLS to force twisted TTLL or TTLL because there is no EO

 

[Twisted101]

Its like using TSLE to force twisted TTLL but my method uses VLS to force twisted TTLL or TTLL because there is no EO

More like OLS than VLS tho

 

[edward_9x]

I've been thinking about ways to reduce from RUF to RU. I've accidentally reinvented CEOR during my experimentation with the concept of 2-gen reduction, so bear with me as I try to figure out something different... Specifically, I've been thinking about another idea:

1. 2x2x2 in DBL
2. 2 gen reduction (EO 2x2x1 then CP)
3. 2 gen finish (<R,U> F2L and 2GLL).

The interesting part is step 2. After solving the 2x2x2 block in DBL, the cube becomes solvable using only <R,U,F>. Therefore, all edges can also be oriented using only <R,U,F>, and all corners can also be permuted using only <R,U,F>. Perhaps one could plan the CP during inspection (before the 2x2x2), then do EO while solving the 2x2x1 from UFR->DFL, then actually solving the CP (both steps done completely <R,U,F>-gen) and solving the rest with <R,U>.

Whether this is nicer than other CP-first methods which reduce to <R,Rw,U> is debateable. Though note that I don't actually know how to solve 2GR-style CP as of right now, and can only do it ZZ-porky style, so therefore I cannot give my opinion on whether it is even possible/feasible to plan this far during 15s of inspection.

What would make this approach hypothetically better than others is the fact that <R,U,F> and <R,U> are the two nicest movesets in speedsolving (<R,U,F> perhaps being even better than <R,Rw,U>). This method also seems nice because of the rotationless aspect, given that you don't need any awkward B moves during the solve (well, maybe except for the first step).

 

[Twisted101]

I've been thinking about ways to reduce from RUF to RU. I've accidentally reinvented CEOR during my experimentation with the concept of 2-gen reduction, so bear with me as I try to figure out something different... Specifically, I've been thinking about another idea:

1. 2x2x2 in DBL
2. 2 gen reduction (EO 2x2x1 then CP)
3. 2 gen finish (<R,U> F2L and 2GLL).

The interesting part is step 2. After solving the 2x2x2 block in DBL, the cube becomes solvable using only <R,U,F>. Therefore, all edges can also be oriented using only <R,U,F>, and all corners can also be permuted using only <R,U,F>. Perhaps one could plan the CP during inspection (before the 2x2x2), then do EO while solving the 2x2x1 from UFR->DFL, then actually solving the CP (both steps done completely <R,U,F>-gen) and solving the rest with <R,U>.

Whether this is nicer than other CP-first methods which reduce to <R,Rw,U> is debateable. Though note that I don't actually know how to solve 2GR-style CP as of right now, and can only do it ZZ-porky style, so therefore I cannot give my opinion on whether it is even possible/feasible to plan this far during 15s of inspection.

What would make this approach hypothetically better than others is the fact that <R,U,F> and <R,U> are the two nicest movesets in speedsolving (<R,U,F> perhaps being even better than <R,Rw,U>). This method also seems nice because of the rotationless aspect, given that you don't need any awkward B moves during the solve (well, maybe except for the first step).

This is just the JJ Method from the Method Development competition 2023

 

[JDAW6]

Yeah bro wth this is my method

 

[V Achyuthan]

I thought of this method (or a variant of some other method I don't know) which could be really good for OH.
1 - 123 block as in Roux
2 - rBd block + EO + FR edge (This step can be done intuitively of algorithmically) while also influencing the next step (i.e to keep the DF edge at the top layer)
This can be done in 2 ways
Way 1
2a) DB edge + Partial EO (which must include putting any Oriented edge in DF position)
2b) Create dBR pair and insert it while completing EO
2c) solve FR edge
OR
Way 2
2a) DB edge + EO
2b) Solve dBR pair + FR edge
3 - Solve CP + the DF edge (which due to influencing from the previous step can now only be in UF, UR, UL or UB)
For recognition, position DFR at UFR recognise the CP case and solve CP and DF edge
4 - Solve the dfR block
5 - 2GLL

Also does anyone know how to generate CP algs in batchsolver?

 

[DoAnKhang]

I thought of this method (or a variant of some other method I don't know) which could be really good for OH.
1 - 123 block as in Roux
2 - rBd block + EO + FR edge (This step can be done intuitively of algorithmically) while also influencing the next step (i.e to keep the DF edge at the top layer)
This can be done in 2 ways
Way 1
2a) DB edge + Partial EO (which must include putting any Oriented edge in DF position)
2b) Create dBR pair and insert it while completing EO
2c) solve FR edge
OR
Way 2
2a) DB edge + EO
2b) Solve dBR pair + FR edge
3 - Solve CP + the DF edge (which due to influencing from the previous step can now only be in UF, UR, UL or UB)
For recognition, position DFR at UFR recognise the CP case and solve CP and DF edge
4 - Solve the dfR block
5 - 2GLL

Also does anyone know how to generate CP algs in batchsolver?

I think it's a variant of triangular francisco

 

[V Achyuthan]

I think it's a variant of triangular francisco

Oh ok. At least now I know it is variant of another method

 

[edward_9x]

Yeah bro wth this is my method

oh lol
in which order did you solve EO, CP and the 221?

 

[JDAW6]

oh lol
in which order did you solve EO, CP and the 221?

EO then 221 then CP then Ru gen ll

 

[Christopher Mowla]

Hi everyone.

A redditor (from rcubers) has a 3x3x3 method proposal and would like your feedback/thoughts/criticism, etc. about it before he generates algs for it.

Of course, if you have an rcubers account and can go reply in that thread, that would be best. But he's okay with just reading the comments/posts from here from you.

(When you go to that topic, you will see why he's not posting here... at least not yet.)

 

[ruffleduck]

Hi everyone.

A redditor (from rcubers) has a 3x3x3 method proposal and would like your feedback/thoughts/criticism, etc. about it before he generates algs for it.

Of course, if you have an rcubers account and can go reply in that thread, that would be best. But he's okay with just reading the comments/posts from here from you.

(When you go to that topic, you will see why he's not posting here... at least not yet.)

The standard which all methods naturally have to compare against are CFOP and Roux. Here are my thoughts.

Comparison of each step against Roux:
1. Roux block but Dl. Virtually the same step 1. Ergonomics may be slightly different, but largely inconsequential.
2. "Finish the bottom layer." This seems comparable to Roux SS. (Honestly, it probably is worse since a) there is more restriction as the pieces are of the same layer as the Roux block, and it's two corners + 1 edge vs 2 edges + 1 corner, but let's give it the benefit of the doubt.)
3. "TELO." I briefly looked at some of the algorithms out of curiosity and, I have to say, they look miserable. Roux at this point only needs to complete RB (having completed FB+SS from steps 1 and 2) which is like 0.3s.
4. "ECO." Impressively, I think the algorithms are even WORSE than those of the previous step. Orienting edges of E layer and CORNERS of U layer, while preserving D layer and segregation of E layer pieces and U layer pieces, is highly restrictive and results in truly awful algs. Roux has CMLL at this stage which is one of the best alg bests in cubing.
5. "ECP." At this point I didn't even bother checking the algs. PLL with E layer mispermutated? My intuition tells me it's worse than PLL (such a stretch I know). Roux's L6E is approximately as good as PLL.

Comparison of each step against CFOP:
1. Roux block but Dl. Let's say for sake of argument a CFOP solver starts with this as well.
2. "Finish the bottom layer." A CFOP solver may instead solve FL and BL edges to complete 223 block. (Probably faster to complete 223 block than finish bottom layer but let's say they're equal for sake of argument.)
3. "TELO." Ah yes, the questionable alg set. Let's say the CFOP solver solves bR here. Certainly far faster for CFOP.
4. "ECO." Without a doubt in my mind, this is slower than LS + OLL in CFOP.
5. "ECP." Yeah this is just worse than PLL.

I am not convinced this method is worth spending time to develop.

 

[Flummox]

The standard which all methods naturally have to compare against are CFOP and Roux. Here are my thoughts.

Comparison of each step against Roux:
1. Roux block but Dl. Virtually the same step 1. Ergonomics may be slightly different, but largely inconsequential.
2. "Finish the bottom layer." This seems comparable to Roux SS. (Honestly, it probably is worse since a) there is more restriction as the pieces are of the same layer as the Roux block, and it's two corners + 1 edge vs 2 edges + 1 corner, but let's give it the benefit of the doubt.)
3. "TELO." I briefly looked at some of the algorithms out of curiosity and, I have to say, they look miserable. Roux at this point only needs to complete RB (having completed FB+SS from steps 1 and 2) which is like 0.3s.
4. "ECO." Impressively, I think the algorithms are even WORSE than those of the previous step. Orienting edges of E layer and CORNERS of U layer, while preserving D layer and segregation of E layer pieces and U layer pieces, is highly restrictive and results in truly awful algs. Roux has CMLL at this stage which is one of the best alg bests in cubing.
5. "ECP." At this point I didn't even bother checking the algs. PLL with E layer mispermutated? My intuition tells me it's worse than PLL (such a stretch I know). Roux's L6E is approximately as good as PLL.

Comparison of each step against CFOP:
1. Roux block but Dl. Let's say for sake of argument a CFOP solver starts with this as well.
2. "Finish the bottom layer." A CFOP solver may instead solve FL and BL edges to complete 223 block. (Probably faster to complete 223 block than finish bottom layer but let's say they're equal for sake of argument.)
3. "TELO." Ah yes, the questionable alg set. Let's say the CFOP solver solves bR here. Certainly far faster for CFOP.
4. "ECO." Without a doubt in my mind, this is slower than LS + OLL in CFOP.
5. "ECP." Yeah this is just worse than PLL.

I am not convinced this method is worth spending time to develop.

Thank you for the response! I can try to rework some of it, I have a couple other ideas that might work out. I’ll also look at the some of the algs and see if any of them are ok.

 

[Twisted101]

This

The standard which all methods naturally have to compare against are CFOP and Roux. Here are my thoughts.

Comparison of each step against Roux:
1. Roux block but Dl. Virtually the same step 1. Ergonomics may be slightly different, but largely inconsequential.
2. "Finish the bottom layer." This seems comparable to Roux SS. (Honestly, it probably is worse since a) there is more restriction as the pieces are of the same layer as the Roux block, and it's two corners + 1 edge vs 2 edges + 1 corner, but let's give it the benefit of the doubt.)
3. "TELO." I briefly looked at some of the algorithms out of curiosity and, I have to say, they look miserable. Roux at this point only needs to complete RB (having completed FB+SS from steps 1 and 2) which is like 0.3s.
4. "ECO." Impressively, I think the algorithms are even WORSE than those of the previous step. Orienting edges of E layer and CORNERS of U layer, while preserving D layer and segregation of E layer pieces and U layer pieces, is highly restrictive and results in truly awful algs. Roux has CMLL at this stage which is one of the best alg bests in cubing.
5. "ECP." At this point I didn't even bother checking the algs. PLL with E layer mispermutated? My intuition tells me it's worse than PLL (such a stretch I know). Roux's L6E is approximately as good as PLL.

Comparison of each step against CFOP:
1. Roux block but Dl. Let's say for sake of argument a CFOP solver starts with this as well.
2. "Finish the bottom layer." A CFOP solver may instead solve FL and BL edges to complete 223 block. (Probably faster to complete 223 block than finish bottom layer but let's say they're equal for sake of argument.)
3. "TELO." Ah yes, the questionable alg set. Let's say the CFOP solver solves bR here. Certainly far faster for CFOP.
4. "ECO." Without a doubt in my mind, this is slower than LS + OLL in CFOP.
5. "ECP." Yeah this is just worse than PLL.

I am not convinced this method is worth spending time to develop.

This Method should also be compared to other L2L methods, but it doesn’t seem like it is better than others like L2L5.

 

[Flummox]

This

This Method should also be compared to other L2L methods, but it doesn’t seem like it is better than others like L2L5.

Where can I find other L2L methods, and what’s L2L5?

 

[edward_9x]

1. Solve 2 squares with the corners in DLB and DLF, respectively
2. Place D layer corners in the D layer, do Porky v1 style recog and solve CP with F' U F
3. EODL using <R,U,S>-gen
4. 2 gen F2L
5. 2GLL

 

[Twisted101]

1. Solve 2 squares with the corners in DLB and DLF, respectively
2. Place D layer corners in the D layer, do Porky v1 style recog and solve CP with F' U F
3. EODL using <R,U,S>-gen
4. 2 gen F2L
5. 2GLL

Would be good but due to mid-solve cp I don’t believe so, this method can be compared to other cp methods like CEOR. Where cp is at beginning. CP at beginning is much easier than cp in the middle of the solve as you can put all the cp thinking into inspection instead of the solve. Also there are blind spots in this method specifically the DL and BR edges. Along with EODL being pretty hard to recognize, due to the chances of DL edge to be in the hidden DL spot flipped. In conclusion this is basically just worse CEOR due to mid solve cp and it is worse petrus due to harder mid-solve eo

 

[Twisted101]

I thought of this method (or a variant of some other method I don't know) which could be really good for OH.
1 - 123 block as in Roux
2 - rBd block + EO + FR edge (This step can be done intuitively of algorithmically) while also influencing the next step (i.e to keep the DF edge at the top layer)
This can be done in 2 ways
Way 1
2a) DB edge + Partial EO (which must include putting any Oriented edge in DF position)
2b) Create dBR pair and insert it while completing EO
2c) solve FR edge
OR
Way 2
2a) DB edge + EO
2b) Solve dBR pair + FR edge
3 - Solve CP + the DF edge (which due to influencing from the previous step can now only be in UF, UR, UL or UB)
For recognition, position DFR at UFR recognise the CP case and solve CP and DF edge
4 - Solve the dfR block
5 - 2GLL

Also does anyone know how to generate CP algs in batchsolver?

Worse MI1

 

[Twisted101]

Where can I find other L2L methods, and what’s L2L5?

L2L4 - Speedsolving.com Wiki

L2L5

Note: L2L5 is a temporary name for this method, but is the best that I have at the moment. Could also be called advanced LBL if the L2L doesn't fit. Context LBL (layer by layer) methods are usually reserved for beginners due to their relative simplicity and easily visible progress, as well as di...

docs.google.com

 

[Ryangamer2010]

Criei um método para Rubik cuber e o chamo de ECL.

Eocamada

Colar

L8ep

Conheço o método ECP e pensei nele enquanto estava ali deitado, pensando.

Eolayer, orientação da borda e primeira camada.

Coll, Orientação e permutação de cantos.

L8ep, Permutação de fatia E e camada U.

L8ep, variantes de 1 aparência: Permutação de camadas com um algoritmo.

2 aparências, separação de bordas e permutação de camadas.

Variante de 3 aparências, separação de bordas, fatia E e permutação epll.

E eu criei esse método hoje, então não tenho algoritmos e não sei quantos casos esse método tem.